A virtual 3000 leads the top stories of 2012
2012 marks 3000 flights of Linux penguins

2012 top losses: Itanium's future, HPQ value

By Ron Seybold

Second in a series

ReporterNotebookDuring 2012 the recent legacy of Hewlett-Packard pulled down the company's futures and values openly for the first time. The company's 73 years of business had devolved in full. A lawsuit exposed completely the new wart of borrowing R&D dollars, over a full decade, to boost HP revenues via mergers and buy-ups. The future of competition was mortaged for commodity computing. The same lack of R&D appetite that'd left the HP 3000 out in the cold after acquiring Compaq business computing now showed HP was bereft of enterprise intellectual property. Nowhere did the cupboard look more bare than the tech choice that had dumped its MPE/iX futures: Itanium.

ExpandIt became plain that the VP of the BCS Unix-Itanium unit, Martin Fink, pushed a plan that might have grown HP-UX stronger just as 3000 sites were getting serious about investing in Unix. The decline of HP 3000 support contracts was even noted in a 2010 document, one that tried to prove that moving Unix to x86 would benefit HP -- by way of sparking new Integrity sales and stronger support revenues for the last OS developed by Hewlett-Packard, HP-UX. One that remained utterly tied to a single chip, Itanium -- until the HP Odyssey emerges from development.

Concocted as a replacement for Intel x86 chips in 1992, the processor that powers all HP Unix servers was uncovered as a product reduced to earning support profits for HP, while taking earnings out of its partner Intel's pockets since 2007. Oracle did lose its lawsuit to halt Itanium releases. But the magnum of evidence uncorked by Oracle -- hundreds of emails that spoke an astounding honesty about the final HP-built enterprise tech environment -- overflowed in the press as well as the courtroom.

Damages to HP from the Oracle lawsuit may fall on the database maker, but the wreckage will not be measured by HP's greatest loss: company valuation. HP sloughed off 43 percent of its market cap during 2012, the largest US slide for the year and a loss attributed to failed mergers fueled by R&D cuts and layoffs. The evidence from HP emails and slides in 2012 made its case of losing up to $4 billion yearly in Itanium-related profits -- even while the company knew, and withheld, facts from its own sales regions about the dire futures of the chip family. The BCS unit continued its slide as of the November financial report (see p. 7 of HP's PDF).

I revisited the turning point of HP's 3000 and MPE/iX exit, but written much larger -- hundreds of thousands of servers put at risk because HP didn't control its own intellectual property for chips anymore. Intel would have to be satisfied, or paid off. In 2012 we learned the latter plan was picked by a board that was still fleeing R&D in 2010.

I wrote a host of articles during 2012 to keep driving home points about investments in HP's Unix. Most of the analysis meant to show that the customers who transitioned HP dollars from MPE to Unix were re-investing in a technology no longer growing (like HP's measure of the 3000 in 2001), one that needed hundreds of millions of HP R&D to keep moving forward.

Even the company's new transition strategy, HP Odyssey, admitted the marketplace had stopped investing in Integrity servers. The business earned profits for HP in the same way that HP collected earnings from MPE/iX. Support contracts, which HP called Technical Services (TS) monies in confidential emails and slides, had dyed Itanium ink from red to black. All was revealed in court exhibits, dumped by Oracle and catalogued by an All Things D reporter for the marketplace to see. Business Critical Systems including HP-UX continued a sales slide -- and a lack of R&D contributed to a decline of vision the markets could now see.

In the email and PowerPoint slide court exhibits listed on Scribd as Oracle Itanium Exhibits Chronological, (a dynamite browse), an urgent HP management story spilled out. Itanium was dying, Intel wasn't cutting HP's minimum purchase requirements, and the sales force and customers were being kept in the dark. In 2009 while the company was trumpeting the advent of a new Integrity system line, Business Critical Systems VP Martin Fink explained in one email that his BCS mission wasn't motivating HP salespeople to keep Itanium-Integrity growing. Not even Intel could be persuaded to help HP discount those Integrity-based systems to make attractive margins.

From the regions' viewpoint, what they see is that
• we have a non-competitive chip
• we are delayed by more than a year, and
• we get no funding relief from Intel to help with margins and keep us in deals.

In short, they're not very impressed with the BCS worldwide team's ability to drive Intel. The Itanium situation is one of our most closely guarded secrets, and we have not wanted to let the region/field know about it, since all it would do is give them another reason not to sell.

There was another way forward for Itanium and HP's Unix, a concept that required the vendor to acquire Sun's Solaris Unix.

Blackbird RationaleOr if only, Fink figured in 2010, HP could invest in an x86 Unix at a cost of $487 million. It could keep Unix revenues stable into 2018. HP considered saving its Unix business in another way in 2009; Fink exhorted its board of directors to buy struggling Sun Microsystems. Oracle acted more decisively on that buy-up, one which HP code-named "Blackbird." (See slide above from early 2009.)

Oracle swept in to snatch the HP Unix competitor. It left HP facing a reality of selling a second-tier Unix on chips that it was paying Intel $88 million a year just to keep developing -- all while the TS support profits declined, from 3000s right down to Integrity servers. (Click on 2010 slide below for details on perhaps the last HP management slide to mention the HP 3000 and its revenues.)

TS-Growth projections 2010We learned in 2012 that HP knew it had nothing left in its R&D property cupboard to help Unix. Text from the slide below showed HP knew in 2009 that x86 chips would "fulfill all aspects of RISC within 5 years."

Strategic Rationale - Current Situation

• HP-UX is on a death march due to inevitable Itanium trajectory

• Companion Technical Services attach business declines precipitously but with a longer tail than the product business
-- No replacement for 45% revenue and 60% of GM for the TS business
-- TS Value is tied to HP-UX and we do not have a go-forward

• x86 is on a credible trajectory to fulfill all aspects of RISC within 5 years

• Going forward, HP will not own the software IP stack upon which to build value -- the hardware stack gets commoditized

Fink went on to be named as the head of the storied HP Labs during 2012 -- the least technically-proven and most business-savvy leader the labs have ever had. He became a direct report to HP CEO Meg Whitman, who's been given the same kind of save-the-company assignment that Steve Jobs faced at Apple in 1996. In this Chicago Tribune article, note the junk bond status, a rating where HP's debt paper began to drift toward in the fall of 2012.

Already Apple's debt has prompted both Moody's Investors Service Inc. and Standard & Poor's to rate the company's bonds at levels so low that analysts such as Chicago-based Carol Levenson of the Gimme Credit bond industry newsletter are calling them "junk."

Obviously, a third fiscal quarter with a $500 million cash drain would take the company down to the near-zero mark and thus be catastrophic. This, you will recall, is how one goes into bankruptcy.

The Tribune article added that "the fight is far from over" at Apple. But it took a revival of innovative design, over more than a decade, to elevate Apple to a state so powerful it could release tablets which would erase HP's laptop sales growth. A proposed split of HP, to spin off enterprise computing from those laptops, came out in 2012 reports from analysts like Therese Poletti at Market Watch.